We quantified GSH levels in adults with and without ASD using 1H-MRS spectral editing and an optimised quantification methodology that allowed us to improve GSH detection, overcoming previous 1H-MRS studies’ limitations.
On this measure, differences in oxidative metabolism differences were not evident. Additionally, we observed no effects of biological sex on GSH levels either across or between groups and no relationship between GSH and current autism scores as measured by AQ. We observed an effect of medication in the DMPFC region; medicated autistic participants had higher GSH than non-autistic and non-medicated autistic participants.
Our findings are in line both with previous in vivo (Durieux et al., 2016; Endres et al., 2017; Sapey-Triomphe et al., 2021) and post-mortem (Chauhan et al., 2012) literature regarding frontal and occipital GSH levels in ASD. Thus, taken together, GSH levels in frontal and occipital regions are not altered in autistic adults. However, our findings do not exclude the possibility that there are GSH alterations elsewhere in the autistic brain. For example, two post mortem studies reported lower GSH in temporal and cerebellar regions in ASD (Chauhan et al., 2012; Rose et al., 2012), therefore GSH differences in ASD may well be brain-region specific. The reasons for this may be diverse. For instance, this could reflect complex differences in regional cellular composition and antioxidant enzyme distribution (Huang & Philbert, 1995; Kumar et al., 2018). Hence, future studies should investigate in vivo GSH levels in other regions, especially in the temporal and cerebellar cortex (Chauhan et al., 2012; Rose et al., 2012).
Alternatively, it is also possible that altered redox regulation might be present only in distinct subgroups of autistic individuals. For example, mitochondrial dysfunction has been proposed as a source of increased oxidative stress in ASD (Bjørklund et al., 2020) but within a distinct subgroup (Rossignol & Frye, 2012), or associated with co-occurring epilepsy (Frye, 2015).
Age is also potentially important. While an in vivo study in children with ASD reported no GSH differences in frontal, motor, and thalamic regions(Song et al., 2023, BioRxiv), post-mortem research point to indices of altered brain energy metabolism and oxidative stress in temporal, cerebellar, and frontal brain samples of autistic children that were not detected in adult samples (Chauhan et al., 2011). Both brain and peripheral (blood and urine) markers of increased oxidative stress have also been associated with higher symptom severity in autistic children, in particular those with more challenging clinical presentations (Adams et al., 2009; Damodaran & Arumugam, 2011; Ghezzo et al., 2013). The autistic individuals included in the current and previous 1H-MRS studies of GSH (Durieux et al., 2016; Endres et al., 2017; Sapey-Triomphe et al., 2021) were adults with average range intellectual ability and no comorbid epilepsy, which might partially explain why GSH levels were unaltered in these studies.
There was some evidence for a sub-group effect in this study – autistic people taking medication had higher GSH levels in the DMPFC region than non-medicated and non-autistic individuals. We cannot say whether higher GSH levels in the medicated group were explained by greater biological complexity associated with co-occurring mental health difficulties or a consequence of medication or both. However, sertraline (taken by the 6 out of 9 of the medicated ASD participants in our sample) has been reported to increase GSH levels in the brain of mice (Abdel-Salam et al., 2011). Future studies in larger samples should examine potential impact of medication on oxidative metabolism in ASD.
We did not find any evidence for a sex-effect on GSH levels in either region. Preclinical literature points to females being less susceptible to oxidative stress than males (Borrás et al., 2003). While neurochemical sex effects have been reported already for other metabolites in autistic individuals (Fung et al., 2020; O’Neill et al., 2020), our study indicates that these do not extend to GSH. However, we acknowledge that even our study had relatively low numbers of female participants and larger sample sizes may be needed to detect sex-specific effects in general, and specifically in ASD (Mo et al., 2021).
We observed no correlations between GSH and AQ scores in either region. This is partially in line with a previous study from our group where GSH levels in an overlapping dorsomedial prefrontal region were not correlated with ADOS scores in adults with ASD. Thus, together with the absence of group differences, absence of a relationship between GSH and phenotype provides confidence that altered GSH in frontal and occipital regions are not a feature of the autistic adults recruited into this study.
Limitations
The most common physiological marker for oxidative stress is decreased GSH/GSSG ratio (Maher, 2005; Schafer & Buettner, 2001). However, biological levels of GSSG are below detection limits of 1H-MRS (Satoh & Yoshioka, 2006). Hence, we cannot rule out the possibility that GSSG levels, and consequently GSH/GSSG ratio, were altered in the current autistic cohort. While post-mortem studies of ASD reported decreased GSH/GSSG alongside decreased GSH (Chauhan et al., 2012; Rose et al., 2012), whether this measure alone will also be meaningful in in vivo studies in ASD needs to be further investigated. We included only adults with normal to high intelligence, therefore our results do not generalise to the whole spectrum of autism. As discussed, GSH alterations may be present in cohorts with different cognitive and age profiles. Lastly, as expected, the recruitment of autistic females was more challenging than males and, in this study, only seven autistic females were included, corresponding to a 3:1 male:female ratio in the ASD group. This male:female is typical in ASD (Loomes et al., 2017), and although previous studies often excluded females, we acknowledge that our analysis of a potential sex effect is still underpowered and should be specifically addressed in future studies.